[ 29 ]
III. On the Structure and Affinities of Guynia annulata, Dune, with Remarks upon the
Persistence of Palaeozoic Types of Madreporaria. By P. Martin Duncan, M.B.
Lond., F.B.8., Professor of Geology in King's College, London.
Eeceived March 16,— Read May 4, 1871.
I. Introduction. 29
II. The description of the genus Guynia and the species Guynia annulata ........ 32
III. The affinities of the species with the Palaeozoic Cyaihaxonidce 32
IV. The differentiation from Haplophyllia paradocca, Pourtales, and the position of this
species in the Oyathaxonidm . . 33
Y. The affinities of Conosmilian species with the Stauridce and Cyaihaxonidce 34
VI. A notice of the secondary stony Madreporaria with Palaeozoic affinities 36
VII. Conclusion 38
VIII. Explanation of the Plate 39
I. During their comprehensive study of the Fossil Corals of the Palaeozoic rocks,
MM. Milne-Edwards and Jules Haime were impressed with the necessity of founding
the great section of the Madreporaria called the Eugosa ; they established the section
in 1850*, an d confirmed its differentiation in 1860f . The characters of the Eugosa were
then decided to be as follows :— " In this division, which comprehends simple as well as
compound corals, the septal structures never form six distinct systems . . . and appear to
be referable to four primitive elements. Sometimes this arrangement is evidenced by
the great development of four principal septa, or by the existence of a corresponding
number of depressions which are seen at the bottom of the calicular fossa and which
give a crucial appearance to it. In other instances one depression or one large septum
exists so as to interrupt the perfection of the septal star. Occasionally no groupings or
systems can be distinguished ; and the septa are represented by striations which rise up
on the upper surface of the tabulse, or by endothecal vesicles which may be observed on
the inner side of the wall. The corallites are always distinct and separate from each
other, for they are never united by an independent ccenenchyma. The wall is usually
feebly developed. The visceral chamber is usually occupied by a series of tabulae, or by
vesicular endotheca, which often constitutes the bulk of the corallum. The septa,
although often incomplete, are never porous or spongy, and they are rarely granular,
and never have synapticulse attached to their laminse. The individual corallites multiply
by gemmation, and do not undergo fissiparous division. The reproductive buds usually
* Monograph of the British Fossil Corals. London, 1850, Palseontographieal Society,
f Histoire Naturelle des Coralliaires. Paris, 1860.
30 PKOFESSOR P. M. DUNCAN ON THE STETTCTUEE
grow upon the calice of the parent, whose growth they arrest, and thus a superposition
of generations is induced. In some genera the gemmation is lateral/'
This section of the Madreporaria necessarily included a great number of genera ; and
as they all could be readily distinguished from those of the other great sections, the new
arrangement was adopted by paleontologists.
It was all the more acceptable because the predominant idea of the geologists of those
days was favoured by the assertion of the existence of any definite groups of organisms
which were characteristic of and peculiar to certain geological formations. The Palae-
ozoic series of rocks was supposed to contain the fossil remains of a fauna and flora
which became extinct before the deposition of the Triassic sediments took place, and a
great break in the continuity of life on the earth was believed to have happened. Every
generalization which appeared to favour such hypotheses was usually accepted as correct
without being subjected to searching criticism; and consequently the foundation of the
section Eugosa, in contradistinction to those of the Aporosa and Perforata, was supposed
to necessitate the inference that the Palaeozoic Madreporaria differed most essentially
from the Neozoic.
Thus the distinguished author of 'Siluria' writes: — " One of the most important of these
discoveries, resulting from the labours of Professor Milne-Edwards, and his coadjutor,
M. Jules Haime, appears to be, that the majority, if not all, of the corals of the Silurian
system, and indeed of the whole Palaeozoic era, belong to divisions of the coral tribe
unknown in modern seas : with rare exceptions, these groups became extinct at the close
of the Palaeozoic epoch. If this be established, and the large cup- and star-corals
(Zoantharia rugosa) and the massive Millepores (Z. tabulata) be, as a whole, distinct
in structure from the star-corals and Madrepores of the Secondary and Tertiary rocks
and of existing coral-reefs, we gain a new fact in the history of animal life upon the
globe, which is in harmony with results obtained by the study of the Crustacea, Mol-
lusca, and Fish of the older epochs" ( ft Siluria/ 4th edition, 1867, p. 217). Moreover,
in a note to page 220 of the same work, the restriction of the non-rugose corals to the
Mesozoic and Cainozoic periods is inferred.
Although the Zoantharia tabulata are as numerous in the existing coral-faunas as
they were in the Palaeozoic (and some of the genera are closely allied), the presumed
fact of the restriction of the Eugosa to the Palaeozoic formations tempted many to come
to the erroneous conclusion respecting the break in the continuity of coral life at the
end of the Permian age.
The characteristic nature of the Palaeozoic coral-fauna was, moreover, strengthened in
the minds of some by the able manner in which MM. Milete-Edwakds and Jules Haime
overthrew the old classification of the corals of the Muschelkalk and St. Cassian strata
of the Trias, and proved that they were not of Palaeozoic genera. Strengthened by the
opinions of many geologists respecting the limitation of life, a number of able palaeon-
tologists have persisted in refusing credence to any facts which should prove, if they
were no longer called anomalies, that the Eugosa were not restricted to the Palaeozoic
AND AFFINITIES OP CHTYNIA ANNULATA. 31
age, and that there has not been a break in the succession of coral species by descent
since the first of them appeared in the seas of old. If the supporters of the hypothesis
which restricts the Eugosa to the Palaeozoic rocks had studied the great work of the di-
stinguished French zoophytologists so often mentioned by me, they would have found that
the following words occur therein : — " Le groupe des Zoanthaires rugeux . . . se compose
presqu'entierement d'especes fossiles appartenant aux terrains anciens"*. The exception
alluded to was a most remarkable and striking one, which .was well known to every
geologist of note. LojsrsDALEf had described a common fossil which was discovered by
Fittok in the Lower Greens J of Atherfield: it was a coral with rugose characteristics,
and MM. Milne-Edwards and Jules Haime placed it amongst the Eugosa and named
it Holocystis elegans, Lonsdale, sp. The specimens are abundant, and they evidently
grew and lived in the Neocomian seas. The existence of the species was considered to
have been anomalous ; but it excited much attention amongst those palaeontologists who
were disposed to consider such anomalies as broken links in a great chain of evidence.
Any forms which might connect the Neocomian species with the Palaeozoic were eagerly
sought for, but without success ; and the distinctness of the Palseozoic and Neozoic coral-
faunas (excepting the Zoantharia tabulata, about which much may be said) might still
be generally admitted, had not the results of the explorations of the sea-floor by the
Americans and by the naturalists of the ' Porcupine ' expeditions reopened the question.
Count Pour/tales $ found a coral with rugose characteristics amongst the dredgings
which were obtained from off the floor of the sea, five miles distant from the Florida
reef, in 1868 ; he founded a new genus to receive the interesting form, and described
it specifically as Haplophyllia paradoxa^ Pourtales. Fortunately the living tissues were
examined and described.
Within the present year (1871) I have examined numerous specimens of a coral which
is new to science, and which presents most marked rugose peculiarities. The specimens
were dredged up in the last expedition of the < Porcupine ' from off the Adventure Bank
in the Mediterranean §, and their description forms the most important part of this
The presence of two genera of Eugosa in the existing coral-fauna has led me to
examine the rugose peculiarities of several species of the genus Conosmilia which were
described by me in an essay on the Fossil Corals of the Australian Tertiary Deposits || , and
also to reconsider the evidence offered respecting the descent of many Lower Liassic corals
from Palaeozoic Eugosa, and which was published in 1867 %.
With a view to connect this evidence with the results of the reconsideration of the
Australian species just alluded to and the discovery of the recent Eugosa, I have intro-
* Hist. Fat. des Corall. vol. iii. p. 324. f Quart. Journ. Geol. Soc. vol. v. 1849.
J Contributions to the Pauna of the Gulf-stream at great depths, 2nd series, 1868 (L. P. Potjrtales).
§ Carpenter and Jeffreys " On Deep-sea Kesearches," Proc. Eoyal Soc. vol. xix. pp. 175, 176.
|| Ann. & Mag. Eat. Hist. September 1865, and Quart. Journ. Geol. Soe. February 9, 1870.
*[[ Brit. Poss. Corals, Supplement issued for 1867. Palseontographical Society, London.
32 PEOFESSOR P. M. DUNCAN ON THE STEUCTURE
duced in this paper a notice of the species o£ the Secondary rocks which were known
to depart from the usual hexameral type, and which were described by MM. Milne-
Edwards and Jules Haime* and by M. de FnoMENTELf. This course of proceeding
is necessary in order to show how the rugose type has persisted during the Neozoic ages.
II. Genus Guynia.
The corallum is simple and long. The wall is thick and solid. The septa are well
developed, lamellar, unequal, and are continuous from the base to the calice. There
are four systems of septa, and one primary septum is longer and larger than the others.
The columella is essential, and is attached to the larger septa. There is no endotheca.
The costee are visible on the growth-rings of the outside of the wall. There is an
Species Guynia annulata, sp. nov. Plate I. figs. 1-8.
The corallum is long, cylindrical, and narrow ; it is sometimes curved. The accretion-
ridges are well developed and regular, and are marked with prominent short spinules,
laminae, or granules which correspond with the costae. The epitheca ornaments the
ridges, and is delicate. The costse extend over the whole length of the corallum, and
usually exist as flat bands between the close and rather wavy accretion-ridges.
There are four principal septa, one of which is larger than the others at the calice.
The four secondary septa are often as large as the primary, but the eight tertiary septa
are almost rudimentary. There are four systems of septa, and three cycles in each ; none
are exsert. The columella is stout, cylindrical, deeply seated in the calice, and adherent
to the larger septa. The interseptal loculi are large, and the transverse outline of the
corallum is sometimes rather angular. The length of the perfect corallum probably
f- inch, the breadth ^ inch.
Locality. Adventure Bank in 92 fathoms.
The numerous specimens of this coral are in excellent preservation, and their condition
is that of living forms whose soft parts have been crushed or washed out during the
operation of removal from their usual locality. Many of the corals adhered by their
sides to mollusca, and resembled annelid-tubes marked with a regular series of ring-like
III. The numerous growth-rings or accretion-ridges give the species a very palaeozoic
facies, especially when there is a very decided constriction between two annular promi-
nences: this facies is made more decided when the tetrameral arrangement or type of
the septa is noticed and the solid columella is distinguished. The stout wall and the
absence of endotheca are exceptional peculiarities ; but although they are not mentioned
in the diagnosis of the Eugosa by MM. Milne-Edwards and Jules Haime, they are
admitted as characterizing a most important family of them — the Cyathaxonidw.
* Hist. ISTat. des Coralliaires, 1860. t E. de Fromeniel, 'Polypiers fossiles/ 1858-61.
AND AFFINITIES OF GITYMA ANNULATA. 33
The following is the diagnosis of the Cyathaxonidce^ the second family of the section
Bugosa* : —
" Corallum having a well-developed septal apparatus, the laminae extending uninter-
ruptedly from the base to the summit of the visceral chamber, and leaving open fossulse
between them without dissepiments, tabulae, or synapticulse. The primary septa are not
decidedly more developed than the others, and do not form a cross as in most of the
Up to the present time but one genus has been associated with this family, viz. Cya-
thaxonia, Michelinf ; it is thus described by MM. Milne-Edwards and Jules Haime J: —
" The corallum is simple, free, finely pedicellate, and has the shape of an elongate
and curved cone. There is a complete epitheca. There is a well-developed septal
fossula situated on the side of the great curvature. The columella is styliform and
very projecting. The septa are smooth and numerous, and most of them unite with
The accretion-ridges and wall are particularly well marked in Cyathaxonia tortuosa,
Michelin, and the size of the septal fossula varies with the species. The genus was
represented in the Upper-Silurian strata of Gothland, and perhaps in the Ludlow rocks
of England, but its species have not been found in Devonian strata ; nevertheless it is
not a rare fossil genus in the American and Belgian Carboniferous strata. Cyathaoconia
cornu, Michelin, is said to be found in English and Belgian Carboniferous deposits.
The great distinction between Guynia and Cyathaoconia is the absence of the septal
fossula in the first genus ; but its species has a large septum, which is a very marked
rugose peculiarity, and the replacement of such septa by depressions or fossulse is common.
There is therefore no reason why Guynia annulata should not be placed in the family
of the Cyathaxoni&W) and that its genus should not be closely associated with Cyathaxonia^.
IV. Count Pourtales describes the genus Haplophyllia (Plate I. figs. 13-15) as
follows || : —
" Corallum simple, fixed by a broad base, covered with a thick epitheca; columella
styliform, strong, very thick at the base. Interseptal chambers deep, uninterrupted by
tabulae or dissepiments, but filling up solid at the bottom."
An introductory paragraph ^ supplies the defective information respecting the septal
apparatus. He therein states : — " The singular coral next to be described strikes one at
first sight by its resemblance to some of the members of the group of the Rugosa of
Milne-Edwards and Hatme. A closer examination tends to confirm that view, much as
it seems improbable to find a living representative of a group so long extinct. In no
other division of the corals is the septal apparatus subdivided into systems that are mul-
tiples of four ; but such is the case in our specimen, though a little obscured by acci-
* Hist. Nat. des CoraU. vol. iii. p. 329. t Icon. Zooph. 1846.
% Op. eit. p. 329. § I have named the genus after Mr. Gwy^t Jeffreys, P.B.S*
|| Op. cit. p. 140. f Pages 139 and 140.
34 PKOEESSOB P. M. DUNCAN ON THE STEUCTTJEE.
dental causes. Another, though perhaps less important, character is the smoothness of
the septa, which present neither perforations, nor synapticula, nor granulations. Tabulse,
however, there are none, the interseptal characters being open from top to bottom.
Among the Eugosa this character is only found in the family of the Cyathaxonidw, to or
near which, therefore, our coral must find its place. From the genus Cyathaxonia it
differs in being attached by a broad base, and also by the absence of a septal fossula.' ,
Haplophyllia paradoxa, Pourtales.
" Corallum subcylindrical, short, fixed by a broad base ; epitheca thick, wrinkled,
reaching higher than the calice, and forming around it several concentric circles as if
representing the separated borders of several superposed layers. Calice circular, fossa
deep. Septa smooth, without granulations or perforations, not reaching the border of
the calice ; like all the internal parts of the calice, their surface is like enamel. Colu-
mella composed of two smooth conical processes, very thick at the base and tending to
fill up the chambers. Eight septa, larger and connected with the columella, alternating
with smaller ones which touch the columella at a much lower level. A further cycle is
indicated by small ridges of the wall-surface in some of the chambers. No distinction
can be made between primary and secondary septa among the eight larger ones, as they
all appear equal.
" Height about |- inch ; diameter of the calice -J inch.
" The coral was living when obtained ; the polyp was of a greenish colour, but was not
otherwise examined when fresh. After having been in alcohol it could be lifted out entire
from the calice, presenting an exact cast of the chambers. The mouth is surrounded
by a circle of about sixteen rather long tentacles, bluntly tuberculated at the tip.
Outside the circle of tentacles extends a membranous disk with radiating and concen-
This unique specimen was dredged up in 324 fathoms off the Florida reef.
It is evident that this interesting form and that which was dredged off the Adventure
Bank have much in common. Both must be classified amongst the Cyathaxonidce ; and
it is quite possible that future dredgings may discover intermediate forms which will
necessitate the absorption either of the genus Haplophyllia or of Guynia. At present
the shape of the closely allied forms, their septal number, the nature of the columellas,
and the characters of the epithecal structures must be considered to separate them
generically. The large septum, so visible in some of the specimens of Guynia annulata,
constitutes in itself a differentiation.
Admitting the generic alliance to be of the closest, Guynia and Haplophyllia will form
with Cyathaxonia the three genera of the family Cyathaxonidw of the section Eugosa.
V. In describing some fossil corals from the Miocene deposits of Australia in 1865, 1
noticed some species of a new genus in the following manner * : — " The new genus Conos*
* Ann. Mag. Nat. Hist. 1865, xvi. p. 18£.
AND AFFINITIES OF GUYNIA ANNTJLATA. 35
milia possesses the twisted ribbon-shaped columella of the subfamily Caryophyllacece,
the endotheca and septal margin of the Trochosmiliacew, and the irregular septal arrange-
ment which was so common in the corals of the Oolitic age, and which, from its octo-
meral type, reflected the Eugosa of Palaeozoic times."
The Geological Survey of Victoria sent me a great number of Miocene corals for
examination and description, and the species were figured and described in an essay on
the Fossil Corals of the Australian Tertiary Deposits, read before the Geological Society,
February 9, 1870. The four well-marked species of the genus Conosmilia were examined
and reconsidered ; but I could not separate them naturally into two groups, although
three out of the four had the octomeral septal arrangement ; the fourth had the usual
Neozoic hexameral type of septal apparatus. I wrote as follows*: — "The most inter-
esting of the corals from the Cainozoic deposits of South Australia are the Conosmiliw.
It is a genus perfectly Australian in its abnormalities. A simple coral with a pellicular
epitheca, having a beautiful herring-bone ornamentation, with an essential, twisted,
"serialaire" columella with endothecal dissepiments, and with plain septa, which have
the hexameral arrangement in some and the octomeral in others, is a form containing
the elements of several classificatory series. The irregular septal arrangement amongst
the closely allied species may be considered to depend upon atavism. Such octomeral
cyclical arrangements occurred in some genera in the Lower-Greensand period and
during the Oolites, &c."
"When the rugosepeculiarities of three out of thefour species of this genus are considered
in relation with the discoveries of existing corals belonging to the section Eugosa the
opinion that they were due to recurrence to ancestral types may well be modified. Like
Haplojphyllia and Guynia the Conosmiliw did not belong to a reef-fauna, but to those
deep-sea faunas which contain so many persistent types. If the theory that the Conos-
milice were originally of an hexameral septal type is correct, then the three out of the
four known species have departed from it and reflect the peculiarities of the ancient
Eugosa ; but if it be admitted that the genus belonged originally to the tetrameral or
octomeral type (for they are identical), then these three Miocene forms were direct
descendants of the Palaeozoic Eugosa, and the one hexameral species was a modification.
Whichever theory is accepted, the descent from a Palaeozoic type is inferred. There is
an interesting relation between so many recent Australian animals and plants and those
of the late Palaeozoic and early Neozoic ages, that, believing in the possibility of the
persistence of coral types belonging to those remote times, I have investigated the struc-
tures of the Conosmiliw with a view of associating three of the species with the Eugosa.
The result is somewhat remarkable ; for it indicates that if the Conosmiliw can be
regarded asBtcgosa, they must be placed amongst the Stauridw, in the neighbourhood of
the genus Polycoelia, whose species are of Permian age in Europe.
Conosmilia elegans, Dune, Conosmilia lituolus, Dune, and Conosmilia anomala, Dune,*,
have, in addition to the rugose septal arrangement, an endotheca which closes off the
* Quart. Journ. GeoL Soc. vol. xxvi. p. 309.
36 PEOFESSOE P. M. DUNCAN ON THE PERSISTENCE OF
lower portions of the interseptal loculi ; but it is curved and arched, and is dissepimental
rather than horizontal and tabulate. Their fasciculate columella and faint pellicular
epithecas are remarkable structures ; and their costal arrangement, by which the septum
corresponds with the intercostal space, is eminently characteristic of some Rugosa.
They differ from the Cyathaoconidce in having an endotheca; but their completely
lamellar septa and their distinct costse associate them with the next, or rather the first
family of the Rugosa — the Stauridce.
The Stauridce were formed into a family by MM. Milne-Ed waeds and Jules Haime in
1850*, and it was differentiated as follows: —
The septa are well developed, and are formed of perfect laminae, which extend unin-
terruptedly through the length of the visceral chamber ; they are united laterally by
lamellary cross dissepiments, and they are arranged in four systems, usually character-
ized by the presence of four large septa arranged in the shape of a cross. The wall is
well developed and imperforate.
The family contained in 1850 two genera of compound and two of simple corals.
The first are, of course, out of the line of the present communication, except that
one of them, the Eolocystis of the Lower Greensand, offers a remarkable proof of the
persistence of the rugose type.
The second or simple coral genera are Polyccelia and Metriophyllurn.
Polyccelia has no columella, and the dissepimental tissue is in the form of horizontal
tabulae, and in Metriophyllum the septa are grouped in four fasciculi. Had a species
of Polyccelia a fasciculate columella and a few arched dissepiments, it would represent
one of the Conosmilice with the tetrameral type — the Conosmilia lituolus for instance.
The manner in which curved or arched dissepiments are associated with and follow
tabulse in the same rugose corals may be seen in many specimens of Carboniferous
species, so that the distinction between the two conditions is not so great as was thought
formerly. The absence of a columella is a generic distinction.
Conosmilia^ according to the theory of its being a persistent type, should be admitted
into the Stauridce^ in the neighbourhood of the genus Polyccelia.
MM. Milne-Edwards and Jules Haime classify the Stauridce as the first family of
the Rugosa, and the Cyathaxonidce as the second ; and the distinction is the want of
endothecal structures in the last-named natural division.
VI. If the occurrence of a tetrameral septal arrangement in a Miocene genus in which
there is a species with the normal Neozoic hexameral type has any significance with
reference to older forms, corresponding phenomena should be more common in more
ancient faunas, — that is to say, the secondary strata should contain a greater number of
tetrameral and octomeral types combined with the hexameral than the tertiary deposits ;
and the fossil corals of the oldest secondary rocks should retain greater evidences of
their descent from Palaeozoic Eugosa than those of a later date.
The following data may be advanced in proof of the occurrence of these requirements.
* Op. cit. vol. iii. page 324.
PALAEOZOIC TYPES OF MADEEPOEAEIA. 37
The discovery of the xxxgomHolocystis elegans^ Lonsd. sp., in the Necomian has already
been noticed ; it is a species which belongs to the same family as Stauria and Conos-
M. be Feomentel has arranged many genera of Secondary and Tertiary corals according
to their septal types ; and he notices that a doubtful generic title is given to Dimorpho-
ccenia corallina by Etallon, and that the form which belongs to the Middle Oolite coral-
fauna is one of the Rugosa. The other species of the genus, and which is the type of it,
has the hexameral septal arrangement, and is a Neocomian fossil.
The same author notices and describes Pleurostylina corallina from the Middle Oolite,
and proves that, with the normal Neozoic hexameral septal type, it has a relic of the
rugose structure in a large septum which passes into the axial space.
Stephanoccenia is a genus with existing Lower Cretaceous and Middle Oolite species
having the hexameral septal type ; but there are other species found in the Eocene and
in the Lower Chalk which have the octomeral arrangement.
Styloccenia has species with a pentameral type in the Eocene and Lower Cretaceous
deposits, and some with the octomeral septal arrangement in Eocene and Miocene
Stylina has species with the hexameral arrangement in the Upper, Middle, and Inferior
Oolites, and others with the octomeral in the Middle Oolite and Lower Chalk ; moreover
it has species in the Trias and Middle and Inferior Oolites which have the decameral
Cryptoceenia has hexameral types in the Neocomian and in the Middle and Inferior
Oolites; but the Middle and Inferior Oolitic strata contain species of it with the octo-
meral septal arrangement.
Goniocora affords examples of hexameral species in the Upper and Middle Oolites and
in the Lias, whilst there is an octomeral type in the Middle Oolitic rocks.
Astrocoenia has hexameral species in the Eocene and Neocomian deposits, octomeral
in the Lower Cretaceous and Middle and Upper Oolitic strata, and in the Tertiaries of
Castel Gombertof ; but all the species described by me from the lowest Liassic strata
possess the decameral type. The lowest coralliferous secondary deposits of Great
Britain contain badly preserved fossils, and yet the Thecosmilian from the White Lias of
Watchet and the cast of a congeneric form from that of Sparkfield have very rugose
characters %. The Thecosmilice from the " Guinea bed" at Binton (zone of Ammonites
planorbis) have the great septum and thin wall of many Rugosa § ; and the species of
Oppelismilia from the next and higher zone of Ammonites angitlatus has no distinct
septal arrangement, but a thick epitheca and calicular gemmation. The great Astrocce-
nian fauna of the zone is composed of twelve species, all of which have the decameral
septal arrangement, and none of them the hexameral. Many of the Montlivaltice of
the zone are so irregular in their development that they cannot be classified under any
* Reuss, Castel Gomberto, Foss. Anthoz. Kaiser. Akad. der "Wissen. Wien, 1868.
t Ee N uss, op. cit. $ P. M. Duncan, Pal. Soe. Lond. vol. xxi. p. 67. § Id. p. 66 t
38 PEOEBSSOE P. M, DUNCAN ON THE PERSISTENCE OF
type ; others have the hexameral arrangement, and Montlivaltia Murchisonice, Dune,
has its septa collected together in four systems. All the species have epithecate walls.
In the zone of Ammonites JBucMandi the genus Lepidophyllia has a very rugose
facies; and Montlivaltia radiata, Dune, of the zone of Ammonites raricostatus, is
evidently furnished with a septal arrangement on the tetrameral type, the four principal
septa being very large. Even in the Middle Lias, Lepidophyllia hebridensis, Dune, has
a rugose aspect ; and the greatest of all Montlivaltiw^ the Montlivaltia Victoria?, of the
zone of Ammonites Henleyi, has an epithecate wall, although there are six systems of
Thus from the Khaetic beds to the Middle Liassic strata the examples of more or less
modified rugose types are frequent ; for the species with the decameral septal arrange-
ment very probably originated from forms of Kugosa with indefinite septal numbers.
After the age of the Lias to the Tertiary period the septal arrangements of many species
and subgenera appear to be very confused; but still many rugose types persisted, having
the tetrameral disposition or the decameral ; so that if it is admitted (and it may be so
consistently with exact truth) that some of the Triassic corals, especially the Montlivaltiw,
have certain but rather faint rugose characters, there is evidence that there has not been
a marked break in the continuity of coral life.
Doubtless many species have varied and have recurred to their ancestral forms; and
this may account for the appearance of tetrameral or octomeral types late in the world's
history in genera whose older secondary species were of the hexameral type ; but the
persistence of the rugose type, more or less modified, up to the present day can no longer
Probably many genera with hexameral septal arrangements originated in Palaeozoic
times; and I have noticed in a former communication* the interesting relation of the
Carboniferous Heterophyllim and the Devonian Battersbyiw to the corals of the normal
VII. It is very remarkable that the two recent species of Rugosa, Haplophyllia para-
doxal Pourtales, and Guynia annulata, Duncan, should belong to the same family of
the section, and that the tertiary Conosmiliw with Palaeozoic affinities should of necessity
be included in a closely allied family of the Eugosa.
That the American and Mediterranean species should be closely allied is in keeping
with the results of the study of the distribution of deep-sea as well as of shallow-water
forms in those distant localities. The Hippurite limestones of Jamaica contain the same
species of Madreporaria as the Cretaceous rocks of Gosau in Austria ; the dark Eocene
shales of the same island have yielded the same species of Madreporaria as the early
Tertiary deposits of North-western Europe ; the Miocene fauna of the Caribbean islands
contains the characteristic species of the corresponding Falunian deposits of France, Italy,
and Malta ; and even the recent Algae of part of the West-Indian area resemble those
* Philosophical Transactions, 1867, p. 643.
PALAEOZOIC TYPES OF MADEEPORAKIA. 39
of the Mediterranean. As regards the Radiata and the Foraminifera, there has been a
very prolonged correspondence of identical and representative species between the
distant areas, and now the occurrence of closely allied species belonging to the persistent
rugose type attests still further the interesting biological relations between the two
margins of the great Atlantic.
It has been noticed that the Conosmilice of the old Australian seas, now found included
in Midtertiary deposits along the northern shores of Victoria and South Australia,
belong to the Stauridae, and that their close ally in that rugose family is the genus
Polyccelia. This genus is extinct, and formed the characteristic coral-fauna of the very
uncoralliferous Permian deposits. Considering the well-known Triassic, Jurassic, and,
indeed, the Palaeozoic facies of portions of the recent and tertiary Australian faunas,
the establishment of the Conosmilice with their Permian affinities as part of a family of
the Rugosa is highly suggestive.
In conclusion, I think that there can be no doubt about the persistence of the rugose
type of Palaeozoic Madreporaria through the Neozoic formations to the present time, and
that the species with hexameral and decameral septal arrangements descended from
rugose types, and the latter especially from those with an indefinite septal number.
VIII. Explanation op the Plate.
Fig. 1. Portion of the corallum of Guynia annulata fixed to a shell. Magnified.
Fig. 2. A specimen showing the calicular end and the costse. There are eight large
primary septa, and one is united to the columella. Magnified.
Fig. 3. A specimen showing the transverse epithecal markings. Magnified.
Fig. 4. Cross section, magnified. There are eight primary septa and several secondary.
Fig. 5. View of a nearly perfect corallum, showing constrictions, costae, and epitheca.
Figs. 6, 7, 8. Portions of a corallum in which at one end there is an hexameral arrange-
ment of the septa (fig. 7), and midway there is the usual octomeral arrange-
Fig. 9. Corallum of Conosmilia anomala, Dune.
Fig. 10. The calice, magnified, showing eight primary septa.
Fig. 11. Corallum of Conosmilia lituolus y Dune. Magnified.
Fig. 12. The calice, magnified.
. ' (The corallum of Haplojphyllia paradom> Pourtales (from Pourtales's ' Deep
Z? 9 :*\ Sea Corals').
Fig. 15. J
40 PKOFESSOK P. M. DUNCAN ON GUYNIA ANNULATA.
Note.— March 25, 1871.
Some days after this communication was sent to the Royal Society, Mr. J. Gwyist
Jeffkeys, F.R.S., forwarded me several specimens of Guynia annulata which he had
found adherent by their sides to mollusca obtained from the Adventure Bank. These
specimens are well preserved, and one of them shows the small commencement of the
long cylinder of the coral. Others exhibit the columella and the large septa and the
normal septal arrangement (octomeral). But one rather deformed coral exhibited on
a fractured surface which was at right angles to the long axis six large septa instead
of the usual eight : this of course required careful examination and explanation. The
septa in the lower and therefore older part of the coral were clearly irregular in their
growth ; but a section midway between this portion and the fractured part established
the interesting fact that the lower part of the corallum possessed the normal octomeral
septal arrangement, and that the upper, in consequence of the abortion of two septa, had
the Neozoic hexameral type. This is very suggestive in the matter of the evolution of
hexameral from octomeral types, or rather from the tetrameral.
March 24, 1871.
%il. .Tran j, M !
■r -./ ,/- i -I
i vr "VT "TA
[ . "We5Ley:..ferto -Tfith. .
Macii-i-e ir. M.a.r'dor:.'jd<.